SIA: Symbolic Interpretability for Anticipatory Deep Reinforcement Learning in Network Control
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908798480809984 |
|---|---|
| author | Jabbari, MohammadErfan Duttagupta, Abhishek Fiandrino, Claudio Bonati, Leonardo D'Oro, Salvatore Polese, Michele Fiore, Marco Melodia, Tommaso |
| author_facet | Jabbari, MohammadErfan Duttagupta, Abhishek Fiandrino, Claudio Bonati, Leonardo D'Oro, Salvatore Polese, Michele Fiore, Marco Melodia, Tommaso |
| contents | Deep reinforcement learning (DRL) promises adaptive control for future mobile networks but conventional agents remain reactive: they act on past and current measurements and cannot leverage short-term forecasts of exogenous KPIs such as bandwidth. Augmenting agents with predictions can overcome this temporal myopia, yet uptake in networking is scarce because forecast-aware agents act as closed-boxes; operators cannot tell whether predictions guide decisions or justify the added complexity. We propose SIA, the first interpreter that exposes in real time how forecast-augmented DRL agents operate. SIA fuses Symbolic AI abstractions with per-KPI Knowledge Graphs to produce explanations, and includes a new Influence Score metric. SIA achieves sub-millisecond speed, over 200x faster than existing XAI methods. We evaluate SIA on three diverse networking use cases, uncovering hidden issues, including temporal misalignment in forecast integration and reward-design biases that trigger counter-productive policies. These insights enable targeted fixes: a redesigned agent achieves a 9% higher average bitrate in video streaming, and SIA's online Action-Refinement module improves RAN-slicing reward by 25% without retraining. By making anticipatory DRL transparent and tunable, SIA lowers the barrier to proactive control in next-generation mobile networks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22044 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SIA: Symbolic Interpretability for Anticipatory Deep Reinforcement Learning in Network Control Jabbari, MohammadErfan Duttagupta, Abhishek Fiandrino, Claudio Bonati, Leonardo D'Oro, Salvatore Polese, Michele Fiore, Marco Melodia, Tommaso Networking and Internet Architecture Artificial Intelligence I.2.6; I.2.4; C.2.1 Deep reinforcement learning (DRL) promises adaptive control for future mobile networks but conventional agents remain reactive: they act on past and current measurements and cannot leverage short-term forecasts of exogenous KPIs such as bandwidth. Augmenting agents with predictions can overcome this temporal myopia, yet uptake in networking is scarce because forecast-aware agents act as closed-boxes; operators cannot tell whether predictions guide decisions or justify the added complexity. We propose SIA, the first interpreter that exposes in real time how forecast-augmented DRL agents operate. SIA fuses Symbolic AI abstractions with per-KPI Knowledge Graphs to produce explanations, and includes a new Influence Score metric. SIA achieves sub-millisecond speed, over 200x faster than existing XAI methods. We evaluate SIA on three diverse networking use cases, uncovering hidden issues, including temporal misalignment in forecast integration and reward-design biases that trigger counter-productive policies. These insights enable targeted fixes: a redesigned agent achieves a 9% higher average bitrate in video streaming, and SIA's online Action-Refinement module improves RAN-slicing reward by 25% without retraining. By making anticipatory DRL transparent and tunable, SIA lowers the barrier to proactive control in next-generation mobile networks. |
| title | SIA: Symbolic Interpretability for Anticipatory Deep Reinforcement Learning in Network Control |
| topic | Networking and Internet Architecture Artificial Intelligence I.2.6; I.2.4; C.2.1 |
| url | https://arxiv.org/abs/2601.22044 |